I think the problem is that Mogensdatter-3 has the wrong parents, according to the long, repetitive and confusing text that was imported from GEDCOM with Mogensdotter-1 in 2011 - incidentally the text is mostly in Norwegian. It discusses two possible sets of parents and it looks to me like the conclusion is that the other set of parents is more probable than the parents attached to Mogensdatter-3. Part of the problem is that the parents attached to Mogensdatter-3 are more "genealogically attractive", since their line goes a lot further back in time, while little is known about the other set.